Wait, where? I wanna read and savour it.

lol they havnt even put the ads in yet

A lot of it hasn't actually been made, though. The AI companies have put in orders for future production. That future capacity can be redirected with a wave of a pen.

Yep the good times are over for now.

And even if, it might turn into a black hole for an agonizingly long time.

Pretty sure MS will just seize the OpenAi data centers and repurpose them for Azure and rent out compute time or use it for XBox streaming. Non of the hardware will reach the open market.

I would be surprised. Quantum computers haven't even been proven to be theoretically useful.

The amount of cooling they need for a simple quantum CPU with 104 qubits. Can you even do anything with it that is useful, you can't even brake encryption because you don't have enough processing power.

The science is advancing, but then as always is the question of, how you scale it to be economical viable, which is a hard problem by it self to solve.

Don't think the market is interested in tech that they can't hype, or have a natural hype cycle.

Which will create a labor boom, to operate the robots, a la Tesla

Challenge there being that seems to have proven elusive. It's not too surprising, but trying to use machine learning for robotics is actually really hard.

Driving is much easier, training data with video, audio, and other sensor input complete with how the human manipulated steering and two pedals.

But direct human interaction with the environment is both much more complicated than three controls and is not instrumented. They are trying to build training data from remote operators, but it turns out we aren't very good at controlling these things remotely anywhere close to acting directly. We are terrible teachers and there's a fraction of the actionable data that other more successful models had to work with.

If an AI sees a video of someone doing something, it can make a similar video, but can't model how that might map to what it would see as unrelated motor and hydraulic operation.

The hurricanes keep getting stronger

OEM license revenue represents a tiny tiny bit of their financials these days. They could just charge nothing for it and business wise no one probably notice much of a difference.

It is foundational to a lot of what they do, but older devices are just as good for their subscription and tie in revenue. Hell I use my work subscription for office from Linux, complete with OneDrive filesystem synchronization. Microsoft gets all their money from my headcount even as I don't even use Windows.

But that capex could bite them hard if revenue falls to follow from it. That's pretty much the only exposure investors care about.

You need to do a custom program if you want to do that. I mean a traditional program where variables are stored properly.

The models have no memory at all, at every question it starts from scratch, so the clients are just "pretending" it has a memory by simply including all previous questions and answers in your last query. You reply "ok", but the model is getting thousands of words with all the history.

Because each question becomes exponentially expensive, at some point it starts to prune old stuff. It either truncates the content (for example the completely useless meta ai chatbot that WhatsApp forced down the throat loses context after 2-3 questions) or it uses the model itself to have a condensed resume of past interactions, but this is how it hallucinates.

Otherwise it will cost like $1 per question and more

Not him haha. Big fan indeed! He is an unclassifiable art man even within the super eclectic cinema landscape that are Korean movies.